Application of classification trees for assessing the degree of damage to conveyor belts under dynamic stress

This article focuses on determining the relationship between the occurrence of severe conveyor belt damage and selected parameters—specifically, the height and weight of the falling material, as well as the type and strength of the conveyor belt itself. To classify and categorize conveyor belt samples into appropriate damage levels, a classification model was developed using a decision tree algorithm. The quality of this model was subsequently validated on a test dataset. Its effectiveness and predictive capability were confirmed through an analysis of overall accuracy, the Kappa coefficient, and the ROC curve, thereby demonstrating its reliability for damage prediction.